Study Five: Assessing the relationship between self-reported RRBs and different

This study assessed the internal consistency and test-retest reliability of the RBQ-2A as well as its correlations with the ARI, DISCO and HADS. The correlations between the different measures of imagination with each other and with the IS, RMB and RSB subscales of the RBQ-2A were also assessed. Finally, this study also aimed to replicate findings from Study Four that IS and a limited pattern of activities is

associated with childhood pretend play, using the imaginative activities and limited pattern of self-chosen activities items from the DISCO along with the RBQ-2A. Participants were seen in two sessions, and took part in several measures of

imagination and RRBs, along with the DISCO, measures of IQ and a measure of mental health.

It was expected that there would be significant associations between the imagination measures and RRBs. It was generally expected that poor performance on imagination tasks would be associated with higher levels of RRBs; however, in the case of originality, this relationship may be reversed. I hypothesised on the basis of

previous research and Study Four that this relationship would be specific to the IS subscale of the RBQ-2A and the limited pattern of self-chosen activities item of the DISCO. It was also expected that the different measures of imagination would be significantly associated with each other. Finally, I expected that the RBQ-2A would be significantly and positively associated with both the ARI and RRB items from the DISCO, along with anxiety symptoms. I also expected that the RBQ-2A would show good test- retest reliability and continue to show good internal consistency.

5.3.1 Method

5.3.1.1 Participants

Participants were recruited in a variety of ways: from the Wales Autism

Research Centre’s Research Recruitment Register; through local charities in the South Wales area; personal contact; and the Cardiff University Noticeboard. Inclusion

140

criteria for the study were: having a clinical diagnosis of any ASD, being aged 18 or older, with the ability to read and write English fluently. The only exclusion criterion was having an IQ below 70 as the study was cognitively demanding. Initially 29 participants were recruited; however, one participant withdrew and another participant was excluded due to having an IQ below 70.

The final sample comprised 27 participants, aged between 19 and 63 at the first time point (M=35.69, SD=12.71), seventeen (63%) of whom identified as male. All participants reported receiving a diagnosis of ASD from a clinician. The majority of participants (N=20) were diagnosed with AS. The remaining participants were

diagnosed with ASD (N=4) or HFA (N=3), and eighteen participants were diagnosed as adults. Twenty-two participants provided evidence of their clinical diagnosis in the form of reports or letters. Twenty-three participants reported whether or not they had a diagnosis of anxiety or depression (N anxiety=5; N depression=2, N both=10, N neither=6). Seven participants were currently attending college or university, eleven participants were employed and twenty-three participants identified as being white.

5.3.1.2 Materials and measures

5.3.1.2.1 Demographics

The following demographic information was collected: date of birth; gender; ASD diagnosis; age at diagnosis; anxiety and/or depression diagnosis if applicable; college/university status; age of leaving education; qualifications; employment status; current job; and ethnicity. Where available, participants were also asked for proof of diagnosis (e.g. diagnostic report, letter from diagnosing clinician).

5.3.1.2.2 RRBs

5.3.1.2.2.1 RBQ-2A

The RBQ-2A as described in Chapters Three and Four and Appendix 3 (page 250). Total score and the subscales (IS, RMB and RSB) from Study Three were used for reliability and validity checks, and the subscales were assessed for their relationship with imagination.

141 5.3.1.2.2.2 Adult Routines Inventory

The ARI; Evans et al., 2017) is the adult self-report version of the Childhood Routines Inventory (Evans et al., 1997). The ARI comprises 55 items covering a wide variety of RRBs, scored on a 5-point scale: Not at all/Never (1); A little/Rarely (2); Somewhat/Sometimes (3); Quite a lot/Often (4); Very much/Always (5). See Appendix 8, pages 265-269 for a list of items. Factor analysis of the ARI results in two factors: Repetitive Sensory-Motor Behaviours/Compulsions (RSMBC), equivalent to RSMB; and Rigidity/Insistence on Sameness (RIS), equivalent to IS.

5.3.1.2.3 Imagination and related constructs (generativity, novelty and flexibility)

Chapter Two (page 75) lists the different measures of imagination used in this thesis and the reasons for selecting them. Generativity/fluency were measured using two verbal fluency tasks, the PFT, and the TTCT. Originality and flexibility were also assessed by the TTCT. Finally, the DISCO includes several items that code for the quality of childhood pretend play.

5.3.1.2.3.1 Measures of generativity only

5.3.1.2.3.1.1 Verbal fluency

As a baseline measure of generativity, participants completed letter and semantic fluency tasks (Benton, 1968; Lezak et al., 2004; Turner, 1999b). In the letter fluency task, participants were first asked to name as many words they could possibly think of beginning with the letter T as a practice. Once they named three correct words, they were asked to name as many different words as they could beginning with the letters F, A and S in one minute. For the semantic fluency task participants were asked to think of words belonging to the category items of clothing as a practice task; once they named three correct words they were then asked to name as many animals and as many foods as they could in one minute without repetitions. Participants’ responses were scored in terms of total correct responses (fluency), excluding non- words, repetitions and responses that did not begin with the correct letter or belong in the given category.

5.3.1.2.3.1.2 Personal future task

This task was adapted from the PFT paradigm developed by MacLeod and colleagues (e.g. MacLeod & Byrne, 1996; MacLeod et al., 1993). For this task,

142

participants were given two minutes to name as many possible future events that may happen to them as they could. They were instructed that the events could be trivial or important, but either definitely going to happen or quite likely to happen. There were two conditions: positive (things that you are looking forward to, in other words, things that you will enjoy) and negative (things that you're worried about or not looking forward to, in other words, things that you would rather not be the case or rather not happen). The exact instructions given to participants are provided in Appendix 9 (page 270). In each condition there were three time periods: next week (including today); next year; and next five to ten years. As with the verbal fluency tasks, participants were scored in terms of total output and total valid responses. Responses were counted as correct if they were a future event, not repetitions of a previous event in any timeframe of a condition, and were the correct emotional valence for the condition.

5.3.1.2.3.2 Measures of generativity, novelty and flexibility

5.3.1.2.3.2.1 Torrance Tests of Creative Thinking subtests

Three subscales from the TTCT (Torrance, 2008) were included to further assess imagination. Two subtests were taken from Figural Form A: Incomplete Figures

and Repeated Figures. For the Incomplete Figures task (IFT), participants were given ten minutes to complete ten meaningless line drawings and give each one a title. For the Repeated Figures task (RFT), participants had ten minutes to create as many drawings as possible from thirty stimuli consisting of a pair of parallel lines, again giving each one a title. The third subtest was taken from Verbal Form A and was Unusual Uses of Cardboard Boxes. In this test, participants were given ten minutes to think of as many new and unusual uses for cardboard boxes as possible and write down their answers.

For the IFT and RFT, participants were scored in terms of fluency (generativity) and originality (novelty). The fluency score comprised the number of drawings

completed and titled, excluding repetitions and uninterpretable responses. This results in two total scores out of 10 and 30 respectively. Only responses that were counted under the fluency score could be coded in terms of originality. To assess originality, participant responses were compared against the most common responses

143

per stimulus derived from population norms by the authors of the TTCT. Each participant response that was included under the common responses was given a score of 0, and all other responses a score of 1. As originality may confound with fluency, the proportion of original responses was calculated by dividing the number of original responses by the fluency score.

For the Unusual Uses of Cardboard Boxes (UUCB) task, fluency and originality were scored in the same way. In the case of fluency, any interpretable response that was not an impossible use (e.g. make a live dog) or a repetition was counted as a valid response. Unlike for the figural TTCT, there was not a strict upper limit in terms of how many uses a participant could generate (although the booklet provided 50 lines). The UUCB task was also scored in terms of flexibility. Each response was given a particular category, and the total number of unique categories generated formed the flexibility score for each participant.

Due to the potential subjectivity of scoring participants in terms of originality and flexibility rather than just fluency, two additional researchers coded participants’ responses on the TTCT. One coded all participants’ responses to the figural TTCT, and the other coded seven participants’ responses to the verbal TTCT.

5.3.1.2.4 DISCO-Abbreviated

The DISCO – Abbreviated, a shortened form of the DISCO, was used in this study for three purposes: firstly, to assess ASD traits and confirm diagnoses in the sample; secondly as an additional measure of imagination; and finally as an additional measure of RRBs. The use of the DISCO in this study is particularly important since much of the work regarding the importance of imagination and its relationship with RRBs in ASD is derived from Wing and Gould’s early observations, which also served as the basis for the DISCO. Therefore this interview provides a rich source of information regarding imagination in ASD, compared to other tools which neglect imagination.

The full DISCO is an effective tool for diagnosing ASD that assesses the frequency, severity and impact of an individual’s behaviours and demonstrates good inter-rater reliability and discriminant validity (Leekam et al., 2002; Maljaars et al., 2012; Nygren et al., 2009; Wing et al., 2002) and has been validated in Swedish

144

(Nygren et al., 2009), as well as showing strong agreement with outputs from the ADI- R and ADOS (Maljaars et al., 2012; Nygren et al., 2009). The items comprising the DISCO-Abbreviated were selected from the full interview on the basis of their predictive validity and ability to discriminate between ASD and non-ASD clinical diagnoses (Carrington, Kent et al., 2014).

As mentioned (page 138), participants were interviewed directly in this study, although they were offered to bring a parent with them. If a participant was unable to recall or did not know the answer to an item, they were asked to consult their parents or a close relative after the interview and inform the research team of the answer if possible. For most DISCO items, participants are given two codes; one for whether they show the behaviour currently and another for whether they have ever shown the behaviour. For the purposes of diagnosis, the ever codes are used. Most DISCO items are coded in the following way: marked problem (0), minor problem (1) or no problem (2), although some items have their own specialised coding systems.

5.3.1.2.4.1 Confirmation of diagnoses

There are several algorithms available for determining whether or not an individual should be diagnosed with ASD. In contrast to the full DISCO, to date the only published algorithm for the DISCO-Abbreviated is based on DSM-5 criteria (Carrington, Kent et al., 2014). DSM-5 criteria have been criticised for not being sensitive or specific enough, particularly in terms of excluding more cognitively able individuals (e.g.

McPartland, Reichow & Volkmar, 2012; Taheri & Perry, 2012), which represents a specific problem for the present study; although the DSM-5 algorithm for the full DISCO has been found to show good sensitivity and specificity across age and ability level (Kent, Carrington et al., 2013). Another issue that may be pertinent for the present sample is the fact that the DSM-5 requires the presence of symptoms before the age of three, and such information is not always possible to obtain for adults depending on the age and health of their parents. Nevertheless, as the only published algorithm for the DISCO-Abbreviated, I chose to implement the DSM-5 algorithm.

In light of the potential issues with the DSM-5 algorithm, I also chose to

implement Wing and Gould’s ASD algorithm. This algorithm requires evidence for the following criteria: social impairment; communication impairment; imagination

145

impairment; and repetitive activities, but does not specify that symptoms have to be present in early childhood. Therefore it represents a useful check of diagnosis that does not rely on information which may not be available to my participants.

5.3.1.2.4.2 Imagination items

Item 38 on the DISCO–Abbreviated is imaginative activities, and for this the interviewer asks participants “Does/Did [name]/you have any pretend play or

imaginative activities?” Unlike most DISCO items, imaginative activities is rated on a six point coding scheme (Table 5.1 below). For adults, this item is coded in terms of the highest level reached by the individual. The first three codes (0-2) represent not having imaginative pretend play in the DISCO DSM-5 algorithm, whereas the last three (3-5) represent having imaginative pretend play. Therefore this item is conceptually similar to the pretend play question asked of participants in the previous two studies.

Table 5-1 Study Five: Scoring guideline for imaginative activities from the DISCO.

No pretend play 0 No play with model toys e.g. no interest in the function of trains, cars and dolls, although A may handle them in the same way as any other objects. 1 Plays with real household equipment using it for its real purpose. No

interest in miniatures e.g. sweeps with real broom, digs with real spade. 2 Holds doll, toy animals as if real, e.g. hugs and kisses toys.

Pretend play 3 Goes through simple sequences of actions with toys as if they are real e.g. pushes toy trains and cars along floor as if real, and makes appropriate noises, or tucks doll in bed.

4 Will pour and give a pretend cup of tea to other person spontaneously (if A only drinks from cup rate 3).

5 Goes through longer sequences of actions with toys e.g. has a doll’s tea party, sets up a garage, road and road bridges for play with toy cars.

Three other measures of childhood pretend play are also relevant. Individuals are coded in terms of delayed pretend play (whether or not pretend play was so delayed as to cause concern), shared pretend play (whether or not pretend play was shared, and if it was shared, whether the child dominated the play), and repetitive pretend play (whether the play was varied and flexible, or repetitive). Delayed pretend play is coded at two levels, whereas shared and repetitive pretend play are

146

coded at three levels corresponding to marked, minor or no problem. Participants who are scored between 0 and 2 on imaginative activities are necessarily delayed in terms of pretend play, and are not scored in terms of repetitive and shared pretend play as they do not show pretend play.

5.3.1.2.4.3 RRB items

There are twenty-three items that are coded under RRBs in the DISCO -

Abbreviated DSM-5 algorithm (Appendix 10, Table 8.9, page 271). However, as higher scores on the RBQ-2A indicate higher levels of RRBs, and lower scores on the DISCO indicate marked difficulties, the coding of RRB items in the DISCO was reversed and recoded in line with RBQ-2A scoring (i.e. marked problem [0] became marked problem [3]). As the RBQ-2A measures current behaviour, the current codes from the DISCO were used as the measure of RRBs. Of particular relevance is the DISCO item limited pattern of self-chosen activities. Limited pattern of self-chosen activities codes whether an individual has a wide range of activities and interests that they

spontaneously engage in, or prefers to engage in repetitive tasks and routines. It is analogous to item 20 (what sort of activity will you choose if you are left to occupy yourself?) of the RBQ-2A and was used here to replicate the finding from Study Four26_.

5.3.1.2.5 Other measures

5.3.1.2.5.1 Hospital Anxiety and Depression Scale

The HADS (Zigmond & Snaith, 1983) was included to assess anxiety for the purpose of assessing construct validity as described in Chapter Two (page 62), and also because performance on the personal future task may be related to both anxiety and depression (page 78). The HADS was originally designed for use with physically ill patients but is now widely used as a measure of mental health symptoms, and demonstrates good reliability and validity (e.g. Bjelland, Dahl, Haug, & Neckelmann, 2002; Mykletun, Stordal, & Dahl, 2001). The HADS comprises 14 items measured on a 4-point Likert scale from 0-3. Seven items relate to anxiety and seven relate to

depression, resulting in a total possible score on each scale of 21; for each scale, the clinical cut-off point is eight.

26 _{This item was used rather than item 20 of the RBQ-2A, so that the two items were drawn from the}

147 5.3.1.2.5.2 Wechsler Abbreviated Scale of Intelligence

FSIQ, VIQ and PIQ were measured using the WASI (Wechsler, 1999), which comprises the Vocabulary, Block Design, Similarities and Matrix Reasoning subtests.

5.3.1.3 Procedure and statistical analyses

This study was approved by the Cardiff University School of Psychology Ethics Committee. Once contact had been established, the participant was sent the

information sheet to read before taking part. The study was conducted in two sessions at one of two locations; the university or a local charity. At the start of the first

session, the participant was offered another chance to read the information sheet. Once the participant had provided written informed consent, they filled in their demographic information. The first session comprised the imagination and creativity tasks (verbal fluency, PFT and the TTCT subtests), followed by a break and then the WASI and questionnaires. Nine participants completed the questionnaires digitally, and seventeen completed paper copies, depending on the available resources at the testing site. The length of time between the two sessions varied depending on the availability of the participant. The second session comprised the DISCO – Abbreviated interview, followed by filling in the RBQ-2A for a second time. Following completion of the second session, the participant was then debriefed. Participants were paid at the end of each session and their travel costs reimbursed.

5.3.1.3.1 Counterbalancing in first session

All participants completed the demographics first. They then completed the imagination and creativity tasks, the order of which was counterbalanced so that half the participants completed the verbal fluency and future thinking task first, and the other half completed the TTCT first. Verbal fluency always came before the PFT, but the conditions of the PFT (positive and negative) were also fully counterbalanced. IFT always came before RFT as they are presented this way in the TTCT; however, whether